The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008
1128
4. COMPARISON & ANALYSIS ON ACCURACY
Seek the area by counting the number of pixel that each class
occupies in land coverage classification map of figure 6 and it
was compared with ratio to total area (Table 2. Ratio I) and area
ratio of land use (Table 3. Ratio II). It was presented that
residential area and road is +1,1% with 8.5% and 9.6%,
cultivate site is +1.2% with 33.2% and 33.4%, forest is -3.5%
with 44.5% and 41.0% and water is +0.9% with 14.8% and
15.7%. In case of cultivate site present the most difference, it is
analyzed to be due to ambiguities in classification with
forest. However, total value of two classes are 76.7% and
74.4% showing relatively small difference of 2.3%. As a result,
classification method applied is determined to be reasonable
presenting the difference for each class between land coverage
classification map and land use map of -3.5% ~ +1.2%.
5. CONCLUSIONS
1. With the pixel based coverage classification of Aster VNIR
image, fair result such as Overall Accuracy of 94.27% and
Kappa Coefficient of 0.8664 was acquired.
2. Difference for each class between land coverage classification
map and land use applying object based classification method
and RX detector is presented to be -3.5% ~ +1.2% thus it is
determined that applied classification method is appropriate.
ACKNOWLEDGEMENT
The authors gratefully acknowledge the generous financial
support received from the Korea Remote Sensing Center
(Public Applications Research of Satellite Data : FN06010).
Also thanks to Mr. Jang MACCA for their support.
REFERENCE
1. Van der Sande, C.J.; de Jong, S.M.; de Roo, A.P.J., A
segmentation and classification approach of IKONOS-2
imagery for land cover mapping to assist flood risk and flood
damage assessment. International Journal of Applied Earth
Observation & Geo-information,Jun,2003, Vol. 4 Issue 3, p217,
13p; DOI: 10.1016/S0303-2434(03)00003-5; {AN 10117970)
2. Friedl, Mark A.; Brodley, Carla E.Maximizing Land Cover
Classification Accuracies Produced by Decision Trees at
Continental to... By: IEEE Transactions on Geoscience
&Remote Sensing, Mar99 Part 2 of 2, Vol. 37 Issue 2, p969, 9p.
2 charts, 4 graphs, 2 maps; {AN 1683704)
3. Chang, Chein-I, and Shao-Shan Chiang,2002. Anomaly
detection and classification for hyperspectral imagery. IEEE
Transactions on Geoscience and Remote Sensing, Vol. 40, No.
6, pp. 1314-1325.
4. Reed I. S., and X. Yu, Adaptive multiple-band CFAR
detection of an optical pattern with unknown spectral
distribution. IEEE Trans. Acoustics, Speech and Signal Proc.
38,pp. 1760-1770,October, 1990.
5. Han, Seunghee, High Resolution 3D Terrain Modeling for
Basis Constitution of Ubiquitous Multi-functional
Administrative City "JCorean Society of Civil
Engineering,2007.6.